What comes from the developer's posts The demand figures in this issue are not our measurements but the developer's own publications. His 19 August post carries the panel screenshot with 60,073 wishlists, the 3,800–4,500 added per day, 185 creator coverages through keymailer, and his own forecast of 65,000–70,000 by launch; the 4 August post covers landing in Steam's personal calendar. Everything else — storefronts, followers, demo player counts and the genre cohort — are our own measurements.
01The short version
This issue is not a post-mortem but a forecast. The Sorting Bureau releases on 21 August, and we are recording three launch scenarios before we know the answer. A prediction written down in advance is the only honest way to test a method: checking a model against days that have already closed is checking it against its own past.
The game arrives at launch with 60,073 wishlists and 3,800–4,500 more per day — almost all of that flow comes from the personal calendar on Steam's front page. The upcoming-releases shelf keeps the game in the top 20 in every country we poll, and its position improves daily.
Our forecast for the peak concurrent player count of the first week is around 1,200, with a pessimistic scenario at 500 and an optimistic one at 2,500. That number is higher than any general-purpose method gives, and section 04 explains in detail why we deliberately depart from them.
The third point is about shape. In this genre the peak almost never lands on release day: three of the four comparable August releases kept growing for another two or three days. We saw the same profile in the month's larger releases. A weak first evening here settles nothing.
02What the game brings to launch
The Sorting Bureau is a cozy sorting game set in a quiet Tokyo workshop: clients bring in boxes of clutter and the player sorts buttons, coins, gears and stamps. No timers, no cost to a mistake. Store tags are Casual, Organizing, Puzzle, Cozy. Apelsinka Games self-publishes; the game supports Windows and macOS and earned full Steam Deck verification before release.
The campaign's decisive event happened in early August: the game landed in the personal calendar on Steam's front page. Several series show it at once. Community followers, which had accumulated 327 people across April through July, more than doubled in the first eighteen days of August.
The curve bends twice. The first bend is mid-May, when the page opened to the public: 11 to 25 followers in a day. The second, far sharper one is early August: 327 people on 31 July and 721 on 18 August. The last observed day added 94 followers — more than all of April.
The upcoming-releases shelf tells the same story on a different scale. The game enters Popular Upcoming on 10 August at #88 and reaches #15 in nine days — in all 41 countries we poll, with no regional skew.
The axis is inverted: higher is better. Both games release on 21 August, both are about tidying and sorting, both climb toward launch. The subject leads the whole stretch and arrives eight places ahead of its rival.
03Three launch scenarios
We record three scenarios for the peak concurrent player count of the first week: pessimistic, median and optimistic. Each has its own curve shape, its own conditions, and its own recognisable predecessor among the August releases.
Note the shape, not only the height: in two scenarios out of three the maximum lands on the third day, not on release day. That is how the genre works — section 05 shows where it comes from.
| Scenario | Week peak | Peak lands | Predecessor |
|---|---|---|---|
| Pessimistic | 350–600 | day 1 | Static Dread: The Submarine — 310 |
| Median | 900–1,700 | day 3 | TV Archive — 1,248, Tidy Up Together — 1,300 |
| Optimistic | 2,000–4,000 | day 4 | Sort Them Ducks — 4,040 |
Pessimistic: 500 — about 30%
This happens if the wishlists gathered by the calendar turn out to be "cold": people clicked after seeing the game on the front page, but were not waiting for it. The launch email then converts poorly, the peak lands on the first evening, and the curve halves each day after. The branch is a serious one: in it both general-purpose methods from section 04 turn out right, and the genre adjustment turns out to be a coincidence across five observations.
This is exactly the outcome the developer himself fears, and he has grounds: he points to Static Dread: The Submarine, which came to launch with a very similar storefront picture and peaked at 310 — we took it apart in the first issue.
Optimistic: 2,500 — about 20%
This happens if the genre wave lifts the game the way it lifted the best of them. August 2026 is the month of a gold rush in tidying and sorting games, and the strongest, Sort Them Ducks, peaked at 4,040 players from a weaker upcoming position and kept growing for four days.
Conditions: a rating above 85% in the first day, staying in New & Trending all week, and no price gap with the genre. Then the curve goes up rather than down, like its shelfmates.
The median scenario — about 50% A peak of about 1,200 players on the third day, roughly 660 on the first evening, about 580 by the end of the week. Positive reviews at 80–86%, wishlists at launch of 66,000–69,000. This is our primary forecast, and the one we will hold ourselves to.
04Where these numbers come from
We counted three ways. Two general-purpose methods give about 430 players, the third — the genre one — about 1,900. We chose the third, and below we explain on what grounds we depart from the first two.
Method one — from the demo. The Sorting Bureau's demo shipped on 22 July and by 17 August peaked at 71 concurrent players, doubling in a fortnight. We collected every "game and its demo" pair among releases from 20 July to 17 August and kept those whose demo is comparable to ours — a peak between 50 and 120 players, 39 of them. The ratio of release peak to demo peak has a median of 6.3 with an interquartile range of 2.2–11.4. For a demo peak of 71 that gives 450 players, corridor 160–810.
The distribution matters more than the median here. The ninth decile of that ratio is 26.7: one game in ten with a small demo gets an almost thirtyfold jump at release. The median describes the typical case, but the right tail is so heavy that "the typical case" is a weak guide here.
Method two — from wishlist-chart position. Every August release that held a chart position between #600 and #1500 before launch — eighteen games. Their median peak is 425 with an interquartile range of 220–1,140. The subject arrives at #925, right in the middle of that band.
Two independent methods converged on 425–450. That would be the answer, if both did not ignore the same thing — genre.
Method three — from shelfmates. Three tidying and sorting games sit in that same rank band, and all three beat the band's median by three to ten times: TV Archive: Tidy Up Together from #788 delivered 1,248, Tidy Up Together from #1296 delivered 1,300, Sort Them Ducks from #1210 delivered 4,040. Two more August genre releases — Shelves and Sorcery and Cleaning Up The Puzzle Gallery — never entered the pre-launch chart at all, so method two predicted nothing for them, and they delivered 1,989 and 1,935. The median of those five is 1,935.
Within the genre the link between position and result is weak: #788 delivered 1,248 while #1210 delivered 4,040, and two games with no chart position at all delivered around 1,950. That is the main reason we give a corridor rather than a point. Static Dread: The Submarine falls out of the group: from the best position it delivered a third of the band median — and it is the only narrative horror game among tidying games. The bottom of the group, Konbini Cleanup at #5189 with 147 players, shows that genre alone saves nobody without a storefront.
The strongest objection against us 126 games shipped in this genre over the month, and only five broke a thousand players. The genre is a lottery where 96% of tickets are blank, and the "genre adjustment" looks at first glance like picking winners after the fact.
The objection is answered by what separates the 121 losers from the subject: they had neither storefront demand nor a place on the upcoming shelf. The right comparison group is not "all tidying games" but "tidying games whose storefront worked", and there five cases out of five delivered at least 1,248. Five is few, but all five point the same way and by a wide margin over the median of their rank band.
Hence the centre of the forecast. We take the genre median — 1,935 — and discount it for the game's one measurable weakness: on the upcoming shelf it sits at #15, while TV Archive reached #2, Sort Them Ducks #6 and Tidy Up Together #8. That is the shelf that works on release day itself, and here it is weaker than all three. On top of that, a direct genre rival ships the same day. A discount of roughly one and a half times gives about 1,200.
05How many copies
We measure the online numbers ourselves; sales we do not. Converting one into the other is unreliable: the ratio of copies to peak across the published panels runs from 5 to 200 and depends on how long a player stays in the game each day. So copies are counted from wishlists, and the conversion comes from the Steamworks panels developers posted themselves.
| Game | Snapshot | Wishlists on the eve | Copies in 24h | Conversion | Over a week |
|---|---|---|---|---|---|
| Sir, We Have an Orc Problem | reconstructed | 198,575 | 43,812 | 22.1% | 64.5% |
| Meowgic | reconstructed | 159,927 | 14,456 | 9.0% | 33.2% |
| ReStory | reconstructed | 850,583 | 79,229 | 9.3% | 11.8% |
| The House Always Wins | reconstructed | 68,852 | 5,895 | 8.6% | 42.1% |
| Security 51 | measured | 172,214 | 12,198 | 7.1% | 15.3% |
| Static Dread: The Submarine | measured | 67,557 | 4,776 | 7.1% | — |
The median for day one is 8.8%, spread across 7.1–22.1%. Both edges come down to where the demand came from: Orc Problem arrived with a fan audience, Submarine with a storefront one, collected by the calendar. The subject sits on Submarine's side and gathered its wishlists the same way.
The snapshot column matters more than the percentages. Two anchors here are measured, and both land on exactly 7.1%: Submarine's wishlists were taken the day before release, and Security 51 published the count of release-day emails sent — 172,214, the exact size of the queue at launch. Both games arrived on storefront demand, as the subject does. The other four have no pre-release snapshot, and the stock was recovered backwards — the balance on release day plus the copies sold. That backward count ignores wishlists added on release day itself, so it overstates the stock and understates conversion. Meowgic has both variants: 9.9% on the measured snapshot against 9.0% on the recovered one. The bias is real but small, and it points down — the true median is likely closer to 10%.
The weekly figures look inflated until you ask where sales come from after day one. They come from the shelves — New & Trending and the top sellers. Where the week is measured, the shelf picked the game up in three releases out of four and conversion went past 33%. For ReStory the low figure is saturation: with 850 thousand wishlists, almost everything that could buy did so on day one.
| Scenario | Peak online | Day-one conversion | Copies in 24h | Copies in a week |
|---|---|---|---|---|
| Pessimistic | 500 | 7.1% | ~4,800 | 9–10k |
| Median | 1,200 | 9.0% | ~6,100 | 20–25k |
| Optimistic | 2,500 | 12.5% | ~8,400 | 33–38k |
The arithmetic runs from 67,500 wishlists — the middle of the range the developer named himself. The scenarios come from section 03, and what ties them to copies is mechanism rather than arithmetic: the shelf that delivers a high peak also delivers sales after the first evening.
Checking the lower edge The demo sits at #1889 in Steam's daily-active-users chart. On its own release day the released Submarine sat at #1740, and its published panel puts that day at roughly 1,230 people. The demo pulls the same number every day before release, and for Submarine that day produced 4,776 copies. The lower edge of the corridor holds without wishlists.
Converting to money
The store price was not published when the forecast was frozen. We work from an assumption: $9.99 with a 30% launch discount, i.e. $6.99. The average gross per copy sits about a quarter below the US store price — Submarine's 4,776 copies produced $34,503, i.e. $7.22 against a $9.59 price. For $6.99 that gives $5.24 per copy.
| Scenario | Copies in a week | Gross | Net |
|---|---|---|---|
| Pessimistic | 9,500 | $50k | $29k |
| Median | 22,500 | $118k | $68k |
| Optimistic | 35,500 | $186k | $108k |
Net is gross less refunds, chargebacks and taxes (the published panels keep 80–87%) and less Steam's 30% share. About 58% of gross.
Where the estimate breaks first A 30% launch discount is more aggressive than the genre norm of 10–20%. It pulls copies up and gross down, and neither correction is in the table: at $8.99 the same 22,500 copies give $139k gross instead of $118k. The second risk is one-sided. Refunds ran at 1.9% for Submarine and reach 12% for the average indie release across our panels, and complaints about grind are expected here — see section 10. Every point of refunds takes a point off net.
06In this genre the peak arrives late
We took the daily peaks of four August genre releases and normalised each to its own maximum. It changes how the first evening should be read.
Three games out of four showed 48–78% of their weekly maximum on release day and kept growing for another two or three days. The one exception is Scratch the Ticket: peak on day one, down 55% by day three. It is also the only game in the group whose positive-review share sits below half. Curve shape and rating are directly linked here.
For release day this means one thing: you cannot judge by the first evening. If the counter shows 600–700 players by midnight on 21 August, that is the median scenario, not a failure. Failure looks different — it looks like a drop on day two.
07Five forks that decide the outcome
| Fork | What is known | Which way it pushes |
|---|---|---|
| Wishlist quality | 60,073 were gathered mostly in two weeks from the front-page calendar, not over a long campaign | The main downside risk |
| First-day rating | The genre runs 76–87% positive, but Scratch the Ticket fell to 46% | Both directions |
| Price | The successful group sits at $4.49–5.39; the only pricier game is Scratch the Ticket at 46% | Down if it leaves the corridor |
| Same-day rival | Mall: Tidy Up Together also ships 21 August, #23 against #15 | Down |
| Pulling the demo | The demo is set to be removed on release day, even though its player count tripled in two weeks | Down, on the tail |
The first two forks are what set the weights we placed in section 03. Fifty per cent on the median scenario is a bet that the genre wave works the way it worked all August. Thirty on the pessimistic one is the price of the cold-wishlist risk: it is the only thing that can devalue the whole 60,000 base at once. Twenty on the optimistic one is the chance that our discount for the weak upcoming position was unnecessary. The weights are a judgement, not a calculation: five observations do not yield confidence intervals.
The wishlist-quality fork is the most interesting methodologically. The personal calendar shows a game to someone who was not looking for it: such a wishlist converts to a purchase at a lower rate than one earned by a trailer or a streamer. We do not observe sales and will not pretend we can estimate them — but this fork is exactly what separates the 500 scenario from the 2,500 one.
The demo fork runs the other way and is easy to miss. The demo shipped on 22 July and grew the whole time: 21 players at peak in early August, 71 by the seventeenth. The funnel works and keeps growing. Removing it on release day is standard practice, yet here the channel is cut exactly at its maximum.
08Who else ships on 21 August
We count twenty-three games on the upcoming shelf for that date, twenty of them present in every country we poll. The subject is third by position.
| Game | Popular Upcoming | Countries |
|---|---|---|
| Moo Who? | #3 | 42 |
| VHOLUME | #4 | 42 |
| The Sorting Bureau | #15 | 42 |
| CICADAMATA | #16 | 42 |
| Click the Button | #21 | 42 |
| Mall: Tidy Up Together | #23 | 42 |
| The Last Disease Z | #24 | 42 |
| Settler's Domain | #25 | 42 |
Mall: Tidy Up Together is the third game in one series in a month: Mall Together shipped on 26 July and peaked at 831 players, Tidy Up Together on 13 August and peaked at 1,300. The studio ships them on a conveyor, and the next one lands on the same date. They will share the New & Trending shelf, the genre listings, and the attention of the same streamers.
09The recommendation graph
The upcoming shelf is not the only way a game gets found. The other one is other games' store pages: the "coming soon" block and the recommendation carousels. We collect that graph in full and can count how many pages show it. Here the order reverses.
On the upcoming shelf the subject is third among the day's releases. In the recommendation graph it is first, and by double: 2,491 donor pages against 1,462 for the runner-up, 1,310 visible against 634. The games ranked above it on the shelf appear in other people's pages markedly less often.
Composition matters more than count. The largest donors land squarely in its audience: Bongo Cat with a peak of 167,832 players, Servant of the Lake — the subject of our second issue — with 28,441, then VPet-Simulator, Mini Cozy Room: Lo-Fi, Spirit City: Lofi Sessions, Paralives, Slime Rancher 2. That is the core of the cozy, meditative audience. And the placements are high: on many pages the game sits in slots 1–5 out of six in the block — actually on screen, not buried in a collapsed list.
The last row is the median donor, and it is the honest one: half of the 2,491 pages have a player count of one or zero. The split: 1,140 donors with no players at all, 1,179 below a hundred, 152 between a hundred and a thousand, and only 20 pages above a thousand. Reach in the graph means two dozen working shopfronts and a very long tail of dead pages.
Two traps that are easy to fall into here First: the "coming soon" block disappears once a game ships. 1,575 of the game's 2,491 donors come from exactly there, so comparing it with already-released genre neighbours compares different quantities. Second: our own graph series is young — observation started on 4 August and the source coverage kept growing for another week and a half. Pre-launch graph counts for August releases are therefore understated unevenly, and we cannot yet test whether the graph predicts anything.
The discount we applied in section 04 for the weak upcoming position is probably too large: on the second discovery channel the subject leads every rival of its release day. But we will not move the centre of the forecast on that basis: propping an estimate up with an untested predictor is exactly what we criticise the skewed-sample median for. Let it stand as a recorded argument instead: if the peak overshoots the top of the corridor, the graph is the first candidate for the explanation.
10The rating risk
The developer says up front that he expects negative reviews about the grind: "that is the point of the game". The genre data partly reassures him — and one example alarms.
| Game | Positive | Reviews | Retention |
|---|---|---|---|
| Sort Them Ducks | 87% | 535 | 78% |
| Tidy Up Together | 79% | 72 | 77% |
| TV Archive: Tidy Up Together | 76% | 201 | 77% |
| Static Dread: The Submarine | 82% | 142 | 36% |
| Scratch the Ticket | 46% | 173 | 29% |
"Retention" here is the share of a game's own peak on 18 August, the last full day of observation; the games shipped on 13–14 August, so they differ in age by a day. grind by itself does not sink a rating — three games out of five hold 76–87%. But you can fail in this genre, and failure is unrelated to audience size: Scratch the Ticket had a healthy 821 players and 46% positive at the same time.
And second: tidying games hold their audience better than narrative ones. 77–78% of peak on 18 August against 36% for Static Dread. A meditative loop with no ending lasts longer than a story with a finale — an argument against the most pessimistic scenario.
11How we will check ourselves
The forecast is frozen on 19 August 2026, two days before launch. It can only be checked against points after that date. Here is exactly what we recorded.
- First-week peak. The median scenario is 900–1,700 with the maximum on day three. Below the corridor means the wishlist-quality risk is real and the general-purpose methods were right; above it means the genre wave is stronger than the discount we applied.
- Curve shape. We predict release day will deliver 45–65% of the weekly maximum. If release day turns out to be the peak, the forecast missed in its most interesting place.
- Wishlists at launch. 66,000–69,000. The developer independently said 65,000–70,000, and here we almost exactly agree.
- Positive-review share. 80–86% at the end of week one.
- Upcoming position. It already sits at #15 on the snapshot, so we predict the top 12 by its release date.
If the forecast fails, this issue will not disappear or be quietly edited: we have a standing habit of dissecting our own misses, and a failed prediction maps the limits of a method more precisely than a lucky one.
12Limits
- We do not observe sales or revenue. Concurrent players and reviews are a weak proxy for sales, not sales. The estimate in section 05 is built on other people's panels and an assumed price: it is a calculation, not a measurement, and it enters the check with that caveat.
- The wishlist number is someone else's. The 60,073 comes from a screenshot of a panel published by the developer and we cannot verify it independently. The 3,800–4,500 per day is his word too.
- The comparison group is tiny. Five genre releases with storefront support is an enumeration, not a sample. A median of five numbers has no confidence interval, and we do not present it as if it had one. The entire centre of the forecast rests on those five cases: if they turn out to be a coincidence, the right estimate is 430 players, not 1,200.
- The demo multiplier rests on thirty-nine games across four weeks of a single month. August seasonality is not separated from genre in it, and games with demos differ systematically from games without.
- The recommendation-graph series is young. Observation started on 4 August and source coverage kept growing for another week and a half. The graph in section 08 is therefore good for comparing games against each other on one snapshot, but not for testing whether it predicts outcomes.
- The price is unknown. At the moment the forecast was frozen the game's price had not been announced publicly, and price is one of the forks.
Disclaimer. The materials presented on this page are informational and research in nature. They consist of observations from open sources and estimates built on them: samples are small, the method is imperfect, calculations are approximate, and forecasts and interpretations may turn out wrong. We do not observe or estimate sales and revenue. Nothing here is business, financial or investment advice; any decision you make on the basis of these materials is your own risk.
Observations: Steam storefronts across 41 countries, demo player counts, community followers, reviews and the recommendation graph. Data snapshot — 19 August 2026, ~17:40 UTC; storefront ranks are live and have almost certainly moved by the time you read this. Comparison cohorts: 18 August releases holding a wishlist-chart position between #600 and #1500, 39 game-and-demo pairs among releases from 20 July to 17 August, 126 tidying and sorting games over the month. The wishlist count and the daily gain come from the developer's public post and we cannot verify them independently. The game's page — The Sorting Bureau on Steam.
The researches and datasets behind this issue were prepared on the PlayRecon v0.3 platform.
Numbers taken from the developer's posts are linked to their originals and remain the property of their authors. When quoting or reusing this forecast, please link back to the original.